AI & connected assistants

Use your own AI keys, and let an assistant like Claude or ChatGPT draft content for you — always with your permissions and always through the same review flow.

Your own AI keys

By default PullPress's AI helpers run on our keys, and each action counts towards your workspace's monthly budget. A workspace owner or admin can instead add your own keys under Workspace settings → AI usage, so you pay your provider directly and no platform budget applies:

Anthropic key powers the text helpers (translation, change explanations, quality checks, stale-content refresh).
OpenAI key powers image generation and semantic search.

Set one or both — each provider's features switch to your key independently. Keys are encrypted at rest and, for security, are never shown again after you save them.

Two things to know: a key isn't checked against the provider when you save it (a wrong key just fails the next time AI runs, so paste carefully), and your own key doesn't unlock AI on the Free plan — AI features need a paid plan regardless of whose key pays for the calls.

Connect an AI assistant (MCP)

PullPress speaks MCP (the Model Context Protocol), so an assistant such as Claude or ChatGPT can work in your content. The assistant acts as you, with your exact permissions, and everything it does goes through the same review flow as your own edits — it can draft and submit, but a human still approves.

Two ways to connect

As a connector (OAuth). Add PullPress as an MCP connector in an app like Claude.ai and sign in. It appears under Profile → Connected apps, where you can disconnect it anytime. These connections are short-lived and refresh themselves in the background.
With a personal token. Under Profile → API tokens, create a token (give it a name like “Claude on my laptop”), and paste it into your assistant. The endpoint is /api/mcp/mcp with the token as a Bearer credential.

A personal token is shown only onceand works like a password: it doesn't expire and isn't scoped to one site, so store it safely and revokeany you no longer use (or don't recognise) from the same page.

What an assistant can do

Connected assistants get a small, safe set of tools:

List the sites your account can work on.
Read your brand voice guide, collections and fields.
List and read entries.
Save a draft and submit it for review.

An assistant cannot publish directly — there is no publish tool. On sites that use review, a submission always lands as a change request for a person to approve.

One exception worth flagging: on open-publishingsites (where approved-by-default is the whole point), “submit for review” publishes immediately, just as it does for a person. If you rely on that mode, only connect assistants you trust.

How AI-written changes show up

When an assistant drafts a change it can attach provenance — the model, the brief it worked from, the sources it used and a confidence level. Reviewers then see a Drafted by AI panel on the review page, and the same details are recorded as trailers on the Git commit, so the origin of a change is never a mystery — in the app or in your repository history.

MCP itself isn't plan-gated: anyone with access to a site can connect an assistant, because the assistant uses its ownAI, not the platform's. The in-app AI helpers and your own keys are what require a paid plan.